Review
New insights into rhythmic brain
activity from TMS–EEG studies
Gregor Thut1 and Carlo Miniussi2,3
1
Centre for Cognitive Neuroimaging, Department of Psychology, 58 Hillhead Street, Glasgow G12 8QB, UK
Department of Biomedical Sciences and Biotechnology, University of Brescia, Viale Europa 11, 25123, Brescia, Italy
3
Cognitive Neuroscience Section, IRCCS San Giovanni di Dio Fatebenefratelli, Via Pilastroni 4, 25125, Brescia, Italy
2
There is renewed interest in the functional role of oscillatory brain activity in specific frequency bands, investigated in humans through electroencephalography
(EEG) and magnetoencephalography (MEG) recordings.
In parallel, there is a growing body of research on noninvasive direct stimulation of the human brain via repetitive (rhythmic) transcranial magnetic stimulation (TMS),
and on those frequencies that have the strongest behavioural impact. There is, therefore, great potential in
combining these two lines of research to foster knowledge on brain rhythms, in addition to potential therapeutic applications of rhythmic brain stimulation. Here,
we review findings from this rapidly evolving field linking intrinsic brain oscillations to distinct sensory, motor
and cognitive operations. The findings emphasize that
brain rhythms are causally implicated in cognitive functions.
Basic organizing principles of brain rhythms: a brief
introduction
Rhythmic activity is a fundamental property of neural
elements and is organized in complex patterns depending
on the state of the brain (e.g. sleep or awake) and on the
task that is currently being executed. It can be recorded
non-invasively from the scalp of human participants by
electroencephalography (EEG) or magnetoencephalography (MEG) and many of the oscillatory components of
EEG and MEG signals have been attributed a role in a
variety of brain operations, including aspects of perception,
cognition and action. One of the organizing principles of
rhythmic activity is synchronization of oscillations across
neuronal elements. This can occur locally between neurons
within an area or over longer distances between areas
within a wider network (e.g. see Refs [1–3]). Another
characteristic of rhythmic brain activity is its frequency
(for frequency bands used to classify brain oscillations see
Glossary). Brain oscillations range from 0.05 to 600Hz
[4], with slow wave activity being associated with sleep and
faster oscillations with the awake state. Oscillatory
activity in specific frequency bands has, therefore, been
related to distinct functions. In terms of perception, for
instance, g-oscillations (30–100Hz) have been linked to
perceptual grouping and maintenance in visual memory
(e.g. see Ref. [5]). In contrast, the slower u (4–8Hz) and aoscillations (8–14Hz) have been related to interregional,
long-distance interaction for control of lower-level by
higher-order areas [6–8] or for unification of cognitive
operations through phase-coupling with other frequency
bands [9,10], besides more local processes (see later).
Despite recent advances, the functional meaning of
these frequencies and their interaction remains a field of
intense research. Here, we illustrate how the combination
of EEG with rhythmic brain stimulation (i.e. transcranial
magnetic stimulation [TMS] or transcranial alternating
current stimulation [tACS]) (Box 1), which is now feasible
through recent technical developments (e.g. see Refs
[11,12]), holds promise for advancing our understanding
of the brain rhythms underlying cognition.
Analogies between EEG-, MEG- and TMS-research on
rhythmic brain activity: added value of combining these
techniques
There is renewed interest in ongoing oscillatory brain
activity in the absence of any stimulus input or motor
output as an index of the internal state of the brain, and
linked to it, the extent to which features of this ongoing
activity have predictive power for subsequent sensory
experience or cognitive processes (e.g. see Ref. [13]). This
interest in how the internal state of the brain shapes
forthcoming perception and cognition transpires many
fields in neuroscience (for a recent review see Ref. [14]).
Within this framework, accumulating EEG and MEG evidence has allowed for investigations of the relationship
between the spectral content of oscillatory activity immediately before the appearance of a visual stimulus and its
perceptual fate both in man [8,15–17] and monkey [18].
This approach has attracted much interest as it provides a
means to address the question of the extent to which
fluctuations in pre-stimulus (baseline) activity represent
Glossary
Electroencephalography (EEG) and magnetoencephalography (MEG): noninvasive electrophysiological recording techniques that sample ongoing
electrical brain activity, which oscillates at various frequencies, through
electrodes or sensors placed over the whole scalp.
Frequency bands used to classify brain oscillations: d: delta (0.5–4Hz), u: theta
(4–8Hz), a: alpha (8–14Hz), b: beta (14–30Hz), g: gamma (30–100Hz), fast (100–
200Hz) and ultra fast (200–600Hz).
Transcranial magnetic stimulation (TMS) or transcranial alternating current
stimulation (tACS): techniques that allow for non-invasive rhythmic stimulation of the human brain at specific frequencies, such as those that can be
recorded via EEG and MEG.
Corresponding author: Thut, G. ([email protected]).
182
1364-6613/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2009.01.004 Available online 13 March 2009
Review
Box 1. Rhythmic brain stimulation
Transcranial magnetic stimulation (TMS), illustrated in Figure Ia, is
an instrument that can be used to investigate the brain-behaviour
relationship and explore the excitability of different regions of the
brain.
The technique involves delivering a brief, single, high-intensity
magnetic pulse to the head through a coil. This induces electrical
currents in a focal area underneath the coil, which interacts with
ongoing activity in the neural tissue. These brief currents can
transiently influence behaviour by producing excitation or inhibition
of the stimulated cortical area [74]. Technical development has
made available devices that are capable of delivering several pulses
in rapid sequence up to 100Hz (repetitive TMS [rTMS]). rTMS offers
the opportunity to interact even more effectively with cortical
activity. It has been shown that by using rTMS it is possible to
transiently modulate neural excitability, with the net effect depending on the stimulation frequency. Usually, low frequency (1Hz)
results in inhibition, whereas high frequency stimulation can result
in excitatory changes in the stimulated area [19].
A recently introduced technique called transcranial alternating
current stimulation (tACS) (Figure Ib) involves applying weak
electrical currents to the head. The currents generate an electromagnetic field that modulates neuronal activity. Electrical stimulation is delivered with a battery-driven stimulator by means of a large
electrode located on the area of interest and a reference electrode
that is placed over a neutral area. Current research indicates that
tACS entrains specific EEG frequency bands and induces phenomena specifically connected to the functions of the stimulated region
and its oscillatory activities (i.e. tACS can interact with ongoing
rhythmic brain activity in a frequency specific fashion) [24].
Both these types of rhythmic brain stimulation could provide the
basis to interact with or induce local oscillatory activity, serving as a
tool to investigate the role of ongoing brain rhythms at different
frequencies.
Figure I. (a) TMS. When researchers operate a TMS coil near a participant’s
scalp, a powerful (2T) and rapidly (300 ms) changing magnetic field passes
painlessly through skin and bone. Because the strength of the magnetic field
falls off very rapidly with distance from the TMS coil, it can penetrate only a
few centimetres and this means that only superficial areas of the brain are
most effectively stimulated. The induced electric field causes electric current in
nearby neurons, thus stimulating targeted regions of the cortex. (b) tACS. This
method relies on application of alternating currents through an electrode.
Electrical currents are applied constantly at low intensities (1 mA) over a period
of time in the order of seconds, to achieve changes in cortical activity. The
waveform of the stimulation is sinusoidal and different frequencies can be
used during stimulation (up to 250Hz).
noise or carry functional meaning. By linking local, prestimulus spectral content to performance of specific aspects
of a task, the approach can also provide new information on
potential functional specificity of brain rhythms. The find-
Trends in Cognitive Sciences
Vol.13 No.4
ing of such links would be suggestive of rhythmic activity
conditioning perception and cognition, as compared to
merely being a by-product of the underlying mechanisms
(e.g. see Ref. [4]), that is, of pre-stimulus brain activity
causally shaping and not only correlating with upcoming
perception.
Similar to the aforementioned EEG and MEG studies,
there is a long tradition of research on rhythmic brain
stimulation via TMS [19,20], and more recently tACS [21],
to modulate ongoing neuronal activity locally and thus to
condition behaviour. Although stimulation with frequencies such as 5Hz (within u-band), 10Hz (within a-band) or
20Hz (within b-band) seems to have a similar (temporary
excitatory) outcome, these effects have mostly been investigated in the context of clinical use of repeated trains of
rhythmic TMS (e.g. see Ref. [19]). Only a few studies have
explored the effects of potential frequency ‘entrainment’
during or shortly after a single train of rhythmic stimulation, showing frequency-specific effects on perceptual or
cognitive task performance [22–24].
Combining research on brain rhythms by EEG, MEG
and rhythmic brain stimulation is therefore promising in
several ways. On the one hand, the combination can provide a more in-depth picture on how brain oscillations
relate to cognitive activities. In particular, it allows not
only for an anatomofunctional mapping of oscillatory brain
activity but also can provide new information on the nature
of the link between brain rhythms and cognition (causal vs.
correlative). On the other hand, the combination has the
potential to give way to new ideas on how to interact with
neural elements and networks in their own way of communication (i.e. through inducing oscillations) with the
aim of manipulating functions. Here, we review new findings on these aspects that start to emerge from combined
TMS–EEG studies. Because many of these studies have
yielded results on the a-frequency band [22,25–31], which
is the most prominent rhythm in the awake state in terms
of amplitude (strongest over posterior and central scalp
recording sites overlying occipito-parietal and sensorimotor areas), there is an emphasis on this rhythmic activity
throughout this review.
Do specific frequencies reflect specific functions? aoscillations and the regulation of cortical excitability
versus inhibition
Recent EEG and MEG research has identified posterior aoscillations, recorded before a visual event, as a predictor of
the perceptual fate of the stimulus [8,15–17]. More specifically, the amplitude of a-oscillations over occipito-parietal
sites is inversely related to perception of the forthcoming
visual event [8,15–17] (Figure 1a) with enhancement of apower being observed under conditions requiring suppression of task-irrelevant visual information [32–35]. Based
on these findings, posterior a-band changes have been
proposed to have a role in perceptually relevant tuning
of occipito-parietal areas in the anticipation of visual
events (e.g. see Refs [16,17]), serving the regulation of
the incoming flow of information along the dorsal stream.
This is likely to go along with changes in receptivity within
these areas [16], possibly by modulating signal-gain control of the forthcoming visual input [17].
183
Review
Trends in Cognitive Sciences Vol.13 No.4
Figure 1. Recent findings of a (a–c) correlative and (d) causal link between oscillatory a-band activity and behaviour, illustrating the interest for combining transcranial brain
stimulation and EEG-research to advance our understanding of brain rhythms. (a) Magnetoencephalography: participants were presented with near-threshold stimuli and
asked to report their visual percepts, while oscillatory brain activity was recorded. Specific oscillatory signatures before stimulus onset were related to whether the stimulus
was perceived or not (schematic representation). Perceptually relevant signatures were confined to the occipito-parietal recording sites (see map of spectral power
difference) and to the a-frequency band (see spectral power curves). Low pre-stimulus power was predictive of hits, whereas high pre-stimulus power was predictive of
misses. Adapted, with permission, from Ref. [17]. (b) Transcranial stimulation and Electroencephalography I: probing the excitability of the visual cortex via its direct
stimulation through occipital single-pulse TMS (evoking sensations of lights, called phosphenes, without retinal input), while oscillatory activity was simultaneously
recorded (schematic representation). Low a activity at the time of TMS was indicative of a highly excitable visual cortex (in terms of the likelihood of inducing phosphenes,
black line, Phosphene-yes trials), whereas high-a activity was indicative of a less excitable cortex (red line, Phosphene-no trials). The perceptually relevant a changes were
localized to occipital electrodes under the TMS coil opposite to the perceived phosphenes (see map, right-lateralized TMS). Adapted, with permission, from Ref. [25]. (c)
Transcranial stimulation and Electroencephalography II: analogous to (b), but probing whether variability in motor cortex excitability (tested through the size of MEPs
evoked by single-pulse TMS over the motor cortex) is linked to spontaneous a-band fluctuations (schematic representation). Rolandic a-band changes were inversely linked
to motor cortex excitability (low a predicting high excitability and high a low excitability) (see left map), whereas other a components were unrelated (right maps). Adapted,
with permission, from Ref. [27]. (d) Transcranial rhythmic stimulation: direct rhythmic stimulation of the visual cortex (via tACS) evoked phosphenes only when the
184
Review
Recently, TMS–EEG studies have provided further support for this notion. Direct evidence for a link between local
a-oscillations and local neuronal excitability has been
obtained by directly stimulating the human occipital cortex
via single-pulse TMS, while simultaneously recording
EEG. Over posterior recording sites, spontaneous fluctuations of oscillatory activity have been shown to co-vary
with the likelihood of occipital TMS to induce sensations of
light (phosphenes) [25] (Figure 1b), thought to result from
early visual cortex (V1/V2) stimulation [36]. An inverse
relationship between pre-TMS a-power and neuronal
excitability in visual areas was observed. High a-amplitude over posterior sites indexed low likelihood of inducing
phosphenes, and low a-amplitude a high likelihood, both
over trials [25] (Figure 1b) and across participants [26].
In parallel, TMS–EEG studies have assessed the
relationship between EEG fluctuations and the excitability
of the motor cortex [27–29] (the latter derived from the size
of peripheral motor-evoked potentials [MEPs] induced by
TMS over primary motor cortex [M1]). In analogy to the
aforementioned finding on a and TMS-probed visual cortex
excitability, Sauseng et al. [27] revealed an inverse link
between spontaneous fluctuations of a-power recorded
close to the central (rolandic) sulcus and the peripheral
MEP-magnitude (Figure 1c) (corroborating Ref. [28], but
see also Refs [37,38] for contradictory evidence). Instead of
probing for a link between a and cortical excitability in
spontaneous fluctuations, Brignani et al. [29] studied the
co-modulation of oscillatory brain activity and TMS-probed
cortical excitability after administration of a TMS protocol
(1Hz-stimulation over several minutes) that is known to
have an inhibitory impact on the excitability of the targetarea (e.g. see Ref. [19]). Applied over left M1, this protocol
resulted in local (left rolandic) a-power enhancement,
which was correlated with a simultaneous decrease in
TMS-probed M1-excitability (i.e. with the inhibitory
impact of the 1Hz-protocol [29]).
These TMS–EEG studies thus extend research on the arhythm by indicating that local a-amplitude carries information about the momentary (excitability) state of
neurons within the dorsal processing stream. The findings
show that up- and down-regulation of a, possibly involving
top-down control, is likely to condition the cortex for forthcoming perception or action. This has added to a mounting
body of information on the role of a-oscillations in attentional selection. Posterior a-power in retinotopically organized (visual) areas [34,39] of the occipital and parietal
lobes [40] is modulated by deployment of visual attention
to specific positions in space, with a-power being downregulated in the hemisphere opposite to the attended
location [7,16,35,40,41] and up-regulated opposite to the
unattended portion of space [33–35]. Over posterior areas,
the a-variations thus seem to regulate the flow of incoming
information [17,25,26,32,42], whereas over rolandic areas,
modulation of oscillatory a-band activity might in addition
be instrumental for the control of transforming perceptions or acquired sensorimotor memories into action
[43,44].
Trends in Cognitive Sciences
Vol.13 No.4
Topography of brain-rhythms: are a-oscillations
intrinsic to specific brain regions?
If specific frequencies have specific functions, they should
show some degree of topographic specificity. Over the
scalp, the cerebral a-rhythm is predominant over posterior
and rolandic sites (over rolandic sites alongside b-oscillations). It is modulated by sensory input or motor output
and therefore thought to reflect the spontaneous rhythm of
sensory and sensory-motor areas. In line with this view, it
has been found to originate in calcarine, occipito-parietal
and somatosensory cortex [45,46], generated through complex cortico-cortical and thalamo-cortical interactions [47],
although other cortical sources including auditory cortex,
areas of the ventral visual stream [45] and mechanisms of
generation exist (e.g. see Ref. [48]). If topographically
specific, the a-rhythm should be differentially affected
by direct cortical stimulation via TMS, depending on
stimulation sites. The following TMS–EEG findings support this view.
Applying single-pulse TMS to the motor cortex triggers
transient neuronal oscillations in the a- and/or b-frequency bands [30,49,50], with mixed results as to which
band is more reactive (a-band in Ref. [30]; b-band in Ref.
[50]). Importantly, the observed synchronization of
neuronal activity was found to show topographic specificity because it was less strong in response to stimulation
of dorsal premotor cortex than M1-stimulation [50]. It has
been proposed that the induced oscillations emerge
because of a reset of the stimulated area’s spontaneous
rhythm, thereby offering a unique approach to study the
generation of oscillatory activity in the human brain
[30,49,50].
Additional information as to the topographic specificity of rhythmic a (and b) generators can be gained
from post-hoc analysis of those TMS–EEG studies that
investigated the EEG-changes in the minutes that follow
administration of repeated TMS pulses (rTMS). Thus far,
such rTMS-induced aftereffects have been quantified in a
total of 14 studies (Figure 2). A ‘meta’-analysis points
towards a consistent topographical pattern of frequency
changes. Across the 10 studies that stimulated either M1
[29,51–54], V1/V2 [55] or dorsolateral prefrontal cortex
(DLPFC) [56–59], there is a bias for a- and b-oscillations
to be affected more often after sensory or motor cortex
stimulation than after DLPFC-stimulation (n=6/6 studies versus n=1/4 studies). Aftereffects in other frequency bands (d, u) are associated more often with
DLPFC than sensory or motor cortex stimulation (n=3/
4 studies versus n=0/6 studies). This differential distribution of aftereffects as a function of stimulation site and
oscillations reaches statistical significance (according to
a chi-square test: p<0.05), although it might partially be
explained by a focus on the analysis of a and/or b-activity
in M1 studies (but see Refs [51] and [58] for ‘doubledissociations’ of a- versus u-aftereffects after M1- versus
DLPFC-stimulation). Note also that although stimulation with rTMS was focal, the observed effects were
widespread, including many recording sites (Figure 2),
stimulation was applied at specific frequencies, namely at a-rhythmicity in darkness and b-rhytmicity in the light. This suggests frequency-specificity in terms of perception,
and thus causal implication in function. Adapted, with permission, from Ref. [24].
185
Review
Trends in Cognitive Sciences Vol.13 No.4
Figure 2. Aftereffects on brain oscillations following the administration of repeated TMS pulses. Stimulation of early visual (V1/V2) or primary motor cortex (M1) most
frequently affected local oscillatory activity in the a- and b-bands [29,51–55], whereas stimulation of the dorsolateral prefrontal cortex (DLPFC) often interfered with
oscillations in other bands [56–59]. Aftereffects were widespread including effects on parieto-occipital a-oscillations after frontal and parietal stimulation [22], rolandic a and
b-activity after premotor cortex stimulation (PM) [60] and prefrontal g after cerebellar stimulation [63,64], illustrating possible network interactions. Abbreviation: IAF,
individual a frequency.
which can be explained by volume conduction because of
the known coarse topographic resolution of EEG. Yet,
this finding could also have resulted from the rTMS
effect spreading from the target site to functionally connected areas.
Of note is a recent TMS–EEG study that used, for the
first time, two TMS coils to apply synchronous rhythmic
stimulation at a-frequency over two sites [31]. The
study revealed induced interregional coherence in the
a-frequency range between the stimulated hand motor
and visual cortices, as compared to unifocal stimulation
[31].
In conclusion, stimulating the cerebral cortex via TMS
and studying what is triggered in terms of changes in brain
oscillations can further advance our understanding of
brain rhythms and their generation. In terms of a-activity,
TMS–EEG findings so far indicate that a-rhythms can be
entrained by TMS and reveal topographic specificity of
this effect, although no cortical ‘a-map’ is firmly established.
186
Embedding in the cortical network: cortico-cortical
interactions in the modulation of a-activity
If the a-rhythm reflects oscillations of neuronal elements
close to signal input- and output-stages as reviewed earlier, it should be adjustable through top-down control from
higher-order areas involved in attention and movement
control and thus depend on the integrity of these areas.
Here, the combined TMS–EEG approach has provided new
information by virtue of TMS over higher-order areas and
the study of EEG-changes at remote, anatomically connected sites (akin to analogous TMS-studies on changes in
evoked potentials [12]).
Capotosto et al. (NaturePrecedings, doi:10101/
npre.2008.1563.1) have studied the network interactions
involved in posterior a-changes during attention orienting in anticipation of a visual stimulus at pre-cued
positions. TMS was applied over areas known to be
involved in the control of visual spatial attention, namely
frontal eye-field (FEF) and intraparietal sulcus (IPS).
Both FEF- and IPS-stimulation led to a breakdown of
Review
the remote, asymmetric regulation of a-activity contralateral to attended versus unattended space. This
suggests that anticipatory a-rhythms at posterior sites
are controlled by signals from FEF and IPS and that
disruption of this top-down control leads to suboptimal
posterior oscillations.
Cortico-cortical and cerebello-cortical interactions in
the generation of rhythmic activity have also been demonstrated by some of the TMS–EEG studies illustrated in
Figure 2. In line with what has already been mentioned,
changes in posterior a-oscillations have been observed
after either frontal or parietal TMS [22]. TMS to the lateral
premotor cortex has been shown to affect rhythmic aactivity over ipsilateral rolandic, and to a lesser extent,
over contralateral rolandic sites [60]. Because fronto-parietal networks are regulating sensory or motor areas for
attentional selection [61] and motor control [62] through
their functional interconnectivity, these effects on a are
likely to have originated downstream to stimulation.
It is also of interest to note that cerebellar stimulation so
far has been observed to induce aftereffects on g-oscillations at prefrontal sites, probably because of the functional connections from the cerebellum to the prefrontal
cortex [63,64] (Figure 2).
Rhythmic activity: epiphenomenal or causal
manifestations of brain function?
Do brain rhythms have a causal functional role or merely
represent epiphenomenal manifestations of the processes
underlying perception, cognition and action? Historically,
the a-rhythm was considered to reflect a cortical idling
state, and its prominent suppression during visual and
motor tasks [45] a valuable (but epiphenomenal) manifestation of cortical activation. Today, there is a growing body
of evidence for a-enhancement beyond baseline (idling)
levels, speaking against this interpretation [65]. A more
conclusive test is to use transcranial brain stimulation for
rhythmic ‘entrainment’. If, in terms of a-oscillations, prestimulus activity is causally shaping perception, stimulation at a-frequency for a-‘entrainment’ (but not at other
frequencies) should lead to perceptual consequences
during or immediately after stimulation. There is, indeed,
emerging evidence for such a causal link, coming from
studies on the behavioural consequences of rhythmic
stimulation within physiological parameters [22–24].
Rhythmic transcranial brain stimulation has been
shown to selectively modulate perceptual processes as a
function of stimulation frequency. In an innovative study
on a-‘entrainment’, Klimesch et al. [22] showed that frontal
and parietal TMS at individual a-frequency immediately
before execution of a visual spatial task (mental rotation
involving visual imagery and memory) affect task performance relative to stimulation in other frequency-bands.
Using tACS, Kanai et al. [24] have shown that occipital
stimulation most effectively induces phosphenes in darkness when applied at a-frequency, whereas in the light bstimulation was most effective (Figure 1d).
Rhythmic entrainment via transcranial brain stimulation has also proven to selectively affect declarative
memory consolidation. During sleep, oscillatory activity
at low frequencies (<4Hz, slow wave activity) has been
Trends in Cognitive Sciences
Vol.13 No.4
linked to overnight consolidation of memories because local
changes in sleep slow-wave activity occur after day-time
learning, and correlate positively with post-sleep performance improvement [66]. In line with a causal role of sleep
slow-waves in memory consolidation, nocturnal TMS at
<1Hz leads to enhanced slow wave activity in addition to
deepening of sleep [67], and nocturnal tACS used with a
DC-shift at slow wave frequency improves memory consolidation [23] (see Ref. [21] for more details regarding this
stimulation protocol, nominated transcranial sinusoidal
direct current stimulation [tSDCS] in Ref. [21]).
Thus, stimulation at physiologically meaningful rhythms
has domain-specific effects on cognitive activities, which
supports causal implication of cortical rhythms
in cognitive function. Although TMS–EEG work so far
has focused on a subset of brain rhythms and
perceptual or cognitive processes, the various ways of using
the TMS–EEG combination illustrated here are transferable to research on other oscillations and cognitive functions, as long as their generators are accessible via TMS.
Many EEG- and MEG-studies have established links between oscillatory activity and higher cognitive activities
other than those reviewed here. For instance, frontal midline u activity has been identified as a correlate of working
memory maintenance, of episodic memory encoding and
retrieval (for review see Ref. [68]) and of emotional regulation (e.g. see Ref. [69]). Others have linked frontal a-activity
with insight problem solving [70] and specific electrographic
signatures (e.g. long-range g-synchronization) with access of
information to consciousness [71]. A causal link between
many of these rhythmic brain activities and their presumed
functions remains to be elucidated.
Box 2. Questions for future research
The bulk of the reviewed literature has focused on a-power, its link
to perception and action, and modulation through brain stimulation. Taking into account phase-information and post-stimulus
periods in future research is likely to provide further crucial
information on the function of a-rhythmicity [10,65].
It is unknown how the reviewed a-fluctuations relate to other
oscillatory changes in sensory selection, such as those of higher
(g) [18] or lower (d) frequencies [75].
It remains to be seen whether entraining specific oscillations such
as posterior a by rhythmic brain stimulation could be helpful to
promote neurorehabilitation. For example, because hemispatial
neglect patients show asymmetric receptivity of otherwise intact
sensory cortices [61], it is conceivable that these patients would
profit from rhythmic entrainment over sensory areas to bias
perception in desired direction.
Although less prominent, a-generators exist outside occipitoparietal and rolandic sites [45]. Their topographic origin (e.g.
frontal a) or their function (e.g. proposed implication in inhibitory
control [65]) could be specifically tested by either a-resetting
through single-pulse TMS or by rhythmic entrainment protocols
over these areas.
Although rhythmic stimulation at frequencies <30Hz (slow wave,
a and b) modulate performance [22–24], it is unknown whether
the techniques that are available for non-invasive brain stimulation in humans can also entrain g-oscillations in a physiologically
meaningful pattern.
Further empirical work is required to elucidate the role of brain
oscillations in other aspects of higher cognitive functions, such as
working memory, memory encoding and retrieval or consciousness [2,3].
187
Review
Rhythmic entrainment: the interest of biasing function
in desired directions
Could the previously mentioned findings lead to new
implementations of rhythmic TMS as a tool in therapy
and neurorehabilitation? Several investigations indeed
indicate that interacting with cortical activity, by means
of rhythmic transcranial stimulation, can positively influence cognitive performance of patients affected by disorders such as aphasia, unilateral neglect or dementia
[72]. The modification of cortical activity for an adequate
period of time through the use of rhythmic stimulation
might adjust a post-lesionally established, but behaviourally maladaptive pattern of brain oscillation, and thus
provide an opportunity for inducing a new balance within
the affected functional network. The data reviewed here
indicate that by pushing the network towards a new
pattern of activation through restoring adequate synchronization, improvement of sensory and motor disorders or
cognitive functions could be achieved. In line with this
view, it has been demonstrated that neurofeedback training to enhance a-oscillations generated in the auditory
cortex (i.e. the tau-rhythm [45]) can help to reduce auditory
sensations in patients suffering from tinnitus [73], possibly
by locally enhancing inhibitory processes. Future brain
stimulation (TMS or tACS) and EEG research should go
hand-in-hand to further elucidate these and other outstanding questions (Box 2).
Concluding remarks
We have reviewed the evolving field of TMS–EEG research
from which new insights into brain rhythms emerge. We
have shown that the TMS–EEG combination is particularly promising and well suited to address several open
questions on the generation and functional role of brain
oscillations. This field has provided so far novel information mostly on the a-rhythm, namely on its role in
perceptually relevant tuning of occipito-parietal areas,
on its origin of generation in sensory-motor regions and
top-down modulation through higher-order areas. This is
in line with the rapidly progressing understanding of this
frequency’s role in attentional selection. Overall, the findings emphasize that the frequency of neural discharges is
not merely epiphenomenal. Extracting meaning from
ongoing brain oscillations in the intact human brain and
on the potential of rhythmic stimulation to selectively
entrain frequencies in selected patients for restoring brain
function will be a challenging field of research for years to
come.
Acknowledgements
We thank Joachim Gross, Jan-Mathijs Schoffelen and three anonymous
reviewers for comments, and Stefano Bonezzi for preparing Figure I in
Box 1.
References
1 Womelsdorf, T. and Fries, P. (2007) The role of neuronal
synchronization in selective attention. Curr. Opin. Neurobiol. 17,
154–160
2 Tallon-Baudry, C. (2004) Attention and awareness in synchrony.
Trends Cogn. Sci. 8, 523–525
3 Ward, L.M. (2003) Synchronous neural oscillations and cognitive
processes. Trends Cogn. Sci. 7, 553–559
188
Trends in Cognitive Sciences Vol.13 No.4
4 Buzsáki, G. et al. (2004) Neuronal oscillations in cortical networks.
Science 304, 1926–1929
5 Jensen, O. et al. (2007) Human gamma-frequency oscillations
associated with attention and memory. Trends Neurosci. 30, 317–324
6 von Stein, A. et al. (2000) Top-down processing mediated by interareal
synchronization. Proc. Natl. Acad. Sci. U. S. A. 97, 14748–14753
7 Sauseng, P. et al. (2005) A shift of visual spatial attention is selectively
associated with human EEG alpha activity. Eur. J. Neurosci. 22, 2917–
2926
8 Hanslmayr, S. et al. (2007) Prestimulus oscillations predict visual
perception performance between and within subjects. Neuroimage
37, 1465–1473
9 Jensen, O. and Colgin, L.L. (2007) Cross-frequency coupling between
neuronal oscillations. Trends Cogn. Sci. 11, 267–269
10 Palva, S. and Palva, J.M. (2007) New vistas for alpha-frequency band
oscillations. Trends Neurosci. 30, 150–158
11 Komssi, S. and Kähkönen, S. (2006) The novelty value of the combined
use of electroencephalography and transcranial magnetic stimulation
for neuroscience research. Brain Res. Rev. 52, 183–192
12 Taylor, P.C. et al. (2008) Combining TMS and EEG to study cognitive
function and cortico-cortico interactions. Behav. Brain Res. 191, 141–
147
13 Engel, A.K. et al. (2001) Dynamic predictions: oscillations and
synchrony in top-down processing. Nat. Rev. Neurosci. 2, 704–716
14 Silvanto, J. et al. (2008) State-dependency in brain stimulation studies
of perception and cognition. Trends Cogn. Sci. 12, 447–454
15 Ergenoglu, T. et al. (2004) Alpha rhythm of the EEG modulates visual
detection performance in humans. Brain Res. Cogn. Brain Res. 20, 376–
383
16 Thut, G. et al. (2006) Alpha-band electroencephalographic activity over
occipital cortex indexes visuospatial attention bias and predicts visual
target detection. J. Neurosci. 26, 9494–9502
17 van Dijk, H. et al. (2008) Prestimulus oscillatory activity in the
alpha band predicts visual discrimination ability. J. Neurosci. 28,
1816–1823
18 Fries, P. et al. (2001) Modulation of oscillatory neuronal
synchronization by selective visual attention. Science 291, 1560–1563
19 Kobayashi, M. and Pascual-Leone, A. (2003) Transcranial magnetic
stimulation in neurology. Lancet Neurol. 2, 145–156
20 Ridding, M.C. and Rothwell, J.C. (2007) Is there a future for
therapeutic use of transcranial magnetic stimulation? Nat. Rev.
Neurosci. 8, 559–567
21 Antal, A. et al. (2008) Comparatively weak after-effects of transcranial
alternating current stimulation (tACS) on cortical excitability in
humans. Brain Stimulation 1, 97–105
22 Klimesch, W. et al. (2003) Enhancing cognitive performance with
repetitive transcranial magnetic stimulation at human individual
alpha frequency. Eur. J. Neurosci. 17, 1129–1133
23 Marshall, L. et al. (2006) Boosting slow oscillations during sleep
potentiates memory. Nature 444, 610–613
24 Kanai, R. et al. (2008) Frequency dependent electrical stimulation of
visual cortex. Curr. Biol. 18, 1839–1843
25 Romei, V. et al. (2008) Spontaneous fluctuations in posterior alphaBand EEG activity reflect variability in excitability of human visual
areas. Cereb. Cortex 18, 2010–2018
26 Romei, V. et al. (2008) Resting electroencephalogram alpha-power over
posterior sites indexes baseline visual cortex excitability. Neuroreport
19, 203–208
27 Sauseng, P. et al. (2009) Spontaneous locally restricted EEG alpha
activity determines cortical excitability in the motor cortex.
Neuropsychologia 47, 284–288
28 Zarkowski, P. et al. (2006) EEG and the variance of motor evoked
potential amplitude. Clin. EEG Neurosci. 37, 247–251
29 Brignani, D. et al. (2008) Modulation of cortical oscillatory activity
during transcranial magnetic stimulation. Hum. Brain Mapp. 29, 603–
612
30 Fuggetta, G. et al. (2005) Modulation of cortical oscillatory activities
induced by varying single-pulse transcranial magnetic stimulation
intensity over the left primary motor area: a combined EEG and
TMS study. Neuroimage 27, 896–908
31 Plewnia, C. et al. (2008) Enhancement of long-range EEG coherence by
synchronous bifocal transcranial magnetic stimulation. Eur. J.
Neurosci. 27, 1577–1583
Review
32 Jensen, O. et al. (2002) Oscillations in the alpha band (9-12 Hz)
increase with memory load during retention in a short-term
memory task. Cereb. Cortex 12, 877–882
33 Kelly, S.P. et al. (2006) Increases in alpha oscillatory power reflect an
active retinotopic mechanism for distracter suppression during
sustained visuospatial attention. J. Neurophysiol. 95, 3844–3851
34 Rihs, T.A. et al. (2007) Mechanisms of selective inhibition in visual
spatial attention are indexed by alpha-band EEG synchronization.
Eur. J. Neurosci. 25, 603–610
35 Rihs, T.A. et al. (2009) A bias for posterior alpha-band power
suppression
versus
enhancement
during
shifting
versus
maintenance of spatial attention. Neuroimage 44, 190–199
36 Kammer, T. et al. (2005) Transcranial magnetic stimulation in the
visual system. II. Characterization of induced phosphenes and
scotomas. Exp. Brain Res. 160, 129–140
37 Lepage, J.F. et al. (2008) EEG and neuronavigated single-pulse TMS in
the study of the observation/execution matching system: are both
techniques measuring the same process? J. Neurosci. Methods 175,
17–24
38 Mitchell, W.K. et al. (2007) Muscle responses to transcranial
stimulation in man depend on background oscillatory activity. J.
Physiol. 583, 567–579
39 Worden, M.S. et al. (2000) Anticipatory biasing of visuospatial
attention
indexed
by
retinotopically
specific
alpha-band
electroencephalography increases over occipital cortex. J. Neurosci.
20, RC63
40 Siegel, M. et al. (2008) Neuronal synchronization along the dorsal
visual pathway reflects the focus of spatial attention. Neuron 60,
709–719
41 Yamagishi, N. et al. (2008) Attentional changes in pre-stimulus
oscillatory activity within early visual cortex are predictive of
human visual performance. Brain Res. 1197, 115–122
42 Ploner, M. et al. (2006) Oscillatory activity reflects the excitability of
the human somatosensory system. Neuroimage 32, 1231–1236
43 Pineda, J.A. (2005) The functional significance of mu rhythms:
translating ‘‘seeing’’ and ‘‘hearing’’ into ‘‘doing’’. Brain Res. Brain
Res. Rev. 50, 57–68
44 Hummel, F. et al. (2002) Inhibitory control of acquired motor
programmes in the human brain. Brain 125, 404–420
45 Hari, R. et al. (1997) Magnetoencephalographic cortical rhythms. Int. J.
Psychophysiol. 26, 51–62
46 Manshanden, I. et al. (2002) Source localization of MEG sleep spindles
and the relation to sources of alpha band rhythms. Clin. Neurophysiol.
113, 1937–1947
47 Steriade, M. et al. (1990) Basic mechanisms of cerebral rhythmic
activities. Electroencephalogr. Clin. Neurophysiol. 76, 481–508
48 Nunez, P.L. et al. (2001) Spatial-temporal structures of human alpha
rhythms: theory, microcurrent sources, multiscale measurements, and
global binding of local networks. Hum. Brain Mapp. 13, 125–164
49 Paus, T. et al. (2001) Synchronization of neuronal activity in the human
primary motor cortex by transcranial magnetic stimulation: an EEG
study. J. Neurophysiol. 86, 1983–1990
50 Van Der Werf, Y.D. and Paus, T. (2006) The neural response to
transcranial magnetic stimulation of the human motor cortex. I.
Intracortical and cortico-cortical contributions. Exp. Brain Res. 175,
231–245
51 Strens, L.H. et al. (2002) The effects of subthreshold 1 Hz repetitive
TMS on cortico-cortical and interhemispheric coherence. Clin.
Neurophysiol. 113, 1279–1285
52 Oliviero, A. et al. (2003) Persistent effects of high frequency repetitive
TMS on the coupling between motor areas in the human. Exp. Brain
Res. 149, 107–113
Trends in Cognitive Sciences
Vol.13 No.4
53 Tamura, Y. et al. (2005) Functional relationship between human
rolandic oscillations and motor cortical excitability: an MEG study.
Eur. J. Neurosci. 21, 2555–2562
54 Fuggetta, G. et al. (2008) Acute modulation of cortical oscillatory
activities during short trains of high-frequency repetitive
transcranial magnetic stimulation of the human motor cortex: a
combined EEG and TMS study. Hum. Brain Mapp. 29, 1–13
55 Thut, G. et al. (2003) Differential effects of low-frequency rTMS at the
occipital pole on visual-induced alpha desynchronization and visualevoked potentials. Neuroimage 18, 334–347
56 Jing, H. and Takigawa, M. (2000) Observation of EEG coherence after
repetitive transcranial magnetic stimulation. Clin. Neurophysiol. 111,
1620–1631
57 Okamura, H. et al. (2001) EEG modification induced by repetitive
transcranial magnetic stimulation. J. Clin. Neurophysiol. 18, 318–325
58 Schutter, D.J. et al. (2001) Effects of slow rTMS at the right
dorsolateral prefrontal cortex on EEG asymmetry and mood.
Neuroreport 12, 445–447
59 Griskova, I. et al. (2007) The effects of 10 Hz repetitive transcranial
magnetic stimulation on resting EEG power spectrum in healthy
subjects. Neurosci. Lett. 419, 162–167
60 Chen, W.H. et al. (2003) Low-frequency rTMS over lateral
premotor cortex induces lasting changes in regional activation and
functional coupling of cortical motor areas. Clin. Neurophysiol. 114,
1628–1637
61 Vuilleumier, P. and Driver, J. (2007) Modulation of visual processing
by attention and emotion: windows on causal interactions between
human brain regions. Philos. Trans. R. Soc. Lond. B Biol. Sci. 362, 837–
855
62 Rushworth, M.F. et al. (2003) The left parietal and premotor cortices:
motor attention and selection. Neuroimage 20 (Suppl. 1), S89–
S100
63 Schutter, D.J. et al. (2003) High frequency repetitive transcranial
magnetic over the medial cerebellum induces a shift in the
prefrontal electroencephalography gamma spectrum: a pilot study in
humans. Neurosci. Lett. 336, 73–76
64 Pastor, M.A. et al. (2006) Modulation of steady-state auditory evoked
potentials by cerebellar rTMS. Exp. Brain Res. 175, 702–709
65 Klimesch, W. et al. (2007) EEG alpha oscillations: the inhibition-timing
hypothesis. Brain Res. Rev. 53, 63–88
66 Huber, R. et al. (2004) Local sleep and learning. Nature 430, 78–81
67 Massimini, M. et al. (2007) Triggering sleep slow waves by transcranial
magnetic stimulation. Proc. Natl. Acad. Sci. U. S. A. 104, 8496–8501
68 Mitchell, D.J. et al. (2008) Frontal-midline theta from the perspective of
hippocampal ‘‘theta’’. Prog. Neurobiol. 86, 156–185
69 Knyazev, G.G. (2007) Motivation, emotion, and their inhibitory control
mirrored in brain oscillations. Neurosci. Biobehav. Rev. 31, 377–395
70 Kounios, J. et al. (2006) The prepared mind: neural activity prior to
problem presentation predicts subsequent solution by sudden insight.
Psychol. Sci. 17, 882–890
71 Melloni, L. et al. (2007) Synchronization of neural activity across
cortical areas correlates with conscious perception. J. Neurosci. 27,
2858–2865
72 Miniussi, C. et al. (2008) Efficacy of repetitive transcranial magnetic
stimulation/transcranial direct current stimulation in cognitive
neurorehabilitation. Brain Stimulation 1, 326–336
73 Dohrmann, K. et al. (2007) Neurofeedback for treating tinnitus. Prog.
Brain Res. 166, 473–485
74 Walsh, V. and Cowey, A. (2000) Transcranial magnetic stimulation and
cognitive neuroscience. Nat. Rev. Neurosci. 1, 73–79
75 Lakatos, P. et al. (2008) Oscillatory entrainment as a mechanism of
attentional selection. Science 320, 110–113
189